常用扩展

# 一次性安装所有扩展
pg extension install pg_stat_statements,pg_repack,pg_prewarm --auto-restart

# 查找 top 5 慢查询
pg exec "SELECT query, calls, total_exec_time::numeric(10,2) AS ms
         FROM pg_stat_statements ORDER BY total_exec_time DESC LIMIT 5;"

# 重组膨胀表
pg exec -- pg_repack -U admin -d default_db -t my_table --no-superuser-check

# 预热表
pg exec "SELECT pg_prewarm('my_table');"

# UUID 生成（PG 18 内置——无需安装扩展）
pg exec "SELECT gen_random_uuid();"  # v4 随机 UUID（PG 13 起内置）
pg exec "SELECT uuidv7();"           # v7 时间有序 UUID（PG 18 新增）

# uuid-ossp 扩展——仅在需要 v1/v3/v5 UUID 时安装
pg exec "CREATE EXTENSION IF NOT EXISTS \"uuid-ossp\";"
pg exec "SELECT uuid_generate_v1();"  # v1 时间戳 + MAC 地址
pg exec "SELECT uuid_generate_v3(uuid_ns_url(), 'https://example.com');"  # v3 基于 MD5
pg exec "SELECT uuid_generate_v5(uuid_ns_url(), 'https://example.com');"  # v5 基于 SHA-1

# 键值存储
pg exec "SELECT 'theme => dark, lang => en'::hstore -> 'theme';"
# 返回: dark

# 不区分大小写匹配
pg exec "CREATE TABLE users (email citext);
         INSERT INTO users VALUES ('Alice@Example.COM');
         SELECT * FROM users WHERE email = 'alice@example.com';"

# 模糊搜索
pg exec "SELECT name, similarity(name, 'John') FROM users
         ORDER BY similarity(name, 'John') DESC LIMIT 5;"

# 创建三元组索引加速模糊搜索
pg exec "CREATE INDEX users_name_trgm_idx ON users USING gin (name gin_trgm_ops);"

# 去除重音
pg exec "SELECT unaccent('Crème Brûlée');"  -- 返回 'Creme Brulee'

# 组合使用实现不区分重音的模糊搜索
pg exec "SELECT name FROM users WHERE name % unaccent('Creme Brulee');"

# 使用 bcrypt 哈希密码
pg exec "SELECT crypt('my_password', gen_salt('bf'));"

# 验证密码
pg exec "SELECT (crypt('my_password', stored_hash) = stored_hash);"

# 对称加密/解密
pg exec "SELECT pgp_sym_decrypt(pgp_sym_encrypt('secret', 'key'), 'key');"

# 启用特定操作的审计日志
pg exec "SET pgaudit.log = 'read, ddl';"

pg extension install postgis --auto-restart

# 创建包含空间数据的表
pg exec "CREATE TABLE places (id serial, name text, geom geometry(Point, 4326));"
pg exec "INSERT INTO places (name, geom) VALUES
         ('San Francisco', ST_MakePoint(-122.4, 37.8)),
         ('London', ST_MakePoint(-0.1276, 51.5074));"

# 查找 5km 范围内的地点
pg exec "SELECT name FROM places
         WHERE ST_DWithin(geom, ST_MakePoint(-122.4, 37.8)::geography, 5000);"

# 计算两个城市间的距离（米）
pg exec "SELECT ST_Distance(
           ST_MakePoint(-74.006, 40.7128)::geography,
           ST_MakePoint(-0.1276, 51.5074)::geography);"

pg extension install vector --auto-restart

# 创建带向量列的表
pg exec "CREATE TABLE items (id serial PRIMARY KEY, embedding vector(3));"
pg exec "INSERT INTO items (embedding) VALUES ('[1,2,3]'), ('[4,5,6]'), ('[7,8,9]');"

# 查找最近邻（欧几里得距离）
pg exec "SELECT id FROM items ORDER BY embedding <-> '[3,1,2]' LIMIT 5;"

# 创建 IVFFlat 索引加速近似搜索
pg exec "CREATE INDEX items_embedding_idx ON items USING ivfflat (embedding vector_l2_ops);"

pg extension install timescaledb --auto-restart

# 创建超级表
pg exec "CREATE TABLE sensor (time timestamptz NOT NULL, value float);
         SELECT create_hypertable('sensor', 'time');"

# 时间桶聚合
pg exec "SELECT time_bucket('1 hour', time) AS bucket, avg(value)
         FROM sensor GROUP BY bucket ORDER BY bucket;"

pg extension install pg_cron --auto-restart

# 调度每日清理任务（午夜运行）
pg exec "SELECT cron.schedule('cleanup', '0 0 * * *',
         'DELETE FROM logs WHERE created_at < now() - interval ''30 days''');"

# 列出已调度的任务
pg exec "SELECT jobid, schedule, command FROM cron.job;"

# 取消调度
pg exec "SELECT cron.unschedule('cleanup');"

# 创建外部服务器
pg exec "CREATE SERVER remote FOREIGN DATA WRAPPER postgres_fdw
         OPTIONS (host '10.0.0.2', port '5432', dbname 'analytics');"

# 创建用户映射
pg exec "CREATE USER MAPPING FOR current_user SERVER remote
         OPTIONS (user 'reader', password 'secret');"

# 导入远程 schema 作为外部表
pg exec "IMPORT FOREIGN SCHEMA public FROM SERVER remote INTO remote_schema;"

# 像查询本地表一样查询远程数据
pg exec "SELECT * FROM remote_schema.events LIMIT 10;"

pg extension install pg_duckdb --auto-restart

# 在 DuckDB 中执行 SQL —— duckdb.query 是表函数（必须在 FROM 子句中使用）
pg exec "SELECT * FROM duckdb.query(\$\$ SELECT 42 AS answer, 'hello' AS msg \$\$);"

# 将表导出为 Parquet 文件
pg exec "COPY my_table TO '/tmp/my_table.parquet' (FORMAT 'parquet');"

# 直接读取 Parquet/CSV 文件
pg exec "SELECT * FROM duckdb.query(\$\$
         SELECT * FROM read_parquet('/tmp/my_table.parquet') WHERE id > 2 \$\$);"

# 将整个查询下推到 DuckDB 执行（OLAP 加速）
pg exec "SET duckdb.force_execution = true;
         SELECT category, sum(amount) FROM sales GROUP BY category;"

# 纯 DuckDB SQL（完全绕过 PostgreSQL 规划器）
pg exec "SELECT * FROM duckdb.raw_query(\$\$ SELECT range AS n FROM range(1, 6) \$\$);"

# --- 从 S3 / MinIO 读取文件（无需复制文件） ---

# 先创建 S3 secret（use_ssl 是文本类型：'true'/'false'）
pg exec "SELECT duckdb.create_simple_secret('S3', '<access_key>', '<secret_key>',
         '', 'us-east-1', 'path', '', 'minio.example.com:9000',
         's3://bucket', '', 'false');"

# 直接查询 bucket 中的 CSV
pg exec "SELECT * FROM duckdb.query(\$\$
         SELECT category, count(*) AS cnt, sum(sales) AS total_sales
         FROM read_csv('s3://bucket/products.csv', auto_detect=true)
         GROUP BY category ORDER BY total_sales DESC \$\$);"